Graeco-Latina Brunensia 23 / 2018 / 1 https://doi.org/10.5817/GLB2018-1-3 A Multidisciplinary Analysis of Non-Literary Latin Texts from Roman Britain Francesca Cotugno University of Nottingham / University of Oxford Abstract The present paper is focussed on the major corpora of non-literary documents written on tab- ČLÁNKY / ARTICLES lets in Roman Britain. This encompasses the stylus tablets from Londinium-Bloomberg, the ink-written tablets from Carlisle and Vindolanda, and the curse tablets incised on lead. The main purpose of this analysis is to show from a quantitative perspective how these different corpora diverge from each other not only in respect of the writing material but also according to the presence or absence of specific linguistic features: gemination, degemination, vowel syncope, and presence or absence of initial h-. Keywords historical sociolinguistics; Roman Britain; non-literary documents 35 Francesca Cotugno A Multidisciplinary Analysis of Non-Literary Latin Texts from Roman Britain 1. Introduction The scholarly interest in Roman Britain encompasses different fields such as the socio- historical and linguistic (Adams 2007: pp. 581–582; Mullen 2016: pp. 584–585). There is a historical coincidence between Roman rule and the first documents written on tablets – in this case stylus tablets – from the area of Londinium-Bloomberg, as these texts rep- resent the oldest documentation available on writing tablets in Roman Britain (Tomlin 2016: p. xiii). For the present analysis the non-literary sources considered are presented in Table 1. Table 1: Synopsis of the non-literary corpora written on tablet from Roman Britain Corpus Date n. of tablets Londinium-Bloomberg1 50–80 CE 405 ČLÁNKY / ARTICLES Carlisle2 79–105 CE 77 Vindolanda3 85–205 CE 772 Curse tablets4 2nd–4th century CE 200 1234 The Londinium-Bloomberg stylus tablets date to between 50 and 80 CE. Evidence from the texts shows a varied provenance of the writers with the Vangiones, the Nervi, the Lin- gones and also people coming from Noricum mentioned (Tomlin 2016: pp. 51–54). These tablets can be described as generally formal documents since the whole corpus consists mostly of official correspondence (i.e., deeds and contracts). Due to the level of formal- ity of these documents, the non-classical forms available in this corpus are few and they are also not particularly relevant as long as they occur only once. Thus, it is not possible to set a consistent comparison between Classical and non-classical occurrences. The other two corpora, Carlisle and Vindolanda, belonged to the military forts from the area of Hadrian’s wall. The fort of Carlisle, once known as Luguualium, was gar- risoned by the auxiliarii from the Ala Gallorum Sebosiana (Tomlin 1998: pp. 31, 36), My gratitude goes to Dr Béla Adamik (HAS Momentum – ELTE University) and also to Dr Daniela Ur- banová (Masarikova University) which encouraged me to pursue my work further with useful comments and advice. Special thanks also goes to Dr. Alex Mullen for her proofreading and for her comments. This paper was presented at the Second International Workshop on Computational Latin Dialectology, Buda- pest, 30–31 March, 2017. This paper is also part of the LatinNow project (latinnow.eu) whose principal investigator is Dr Alex Mullen. LatinNow aims to situate the phenomena of Latinization, literacy, bi- and multilingualism within broader social developments. The project is hosted by the University of Notting- ham and based at the Centre for the Study of Ancient Documents (CSAD), University of Oxford. LatinNow receives funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 715626). 1 Based on the very recent publication, Tomlin (2016). 2 Tomlin (1998) is the current edition of the Carlisle writing tablets. 3 Composed of the following editions: Bowman & Thomas (1994; 2003) and Bowman, Thomas & Tomlin (2010; 2011). 4 Composed mainly of Tomlin (1988; 1993) concerning the main sites of Bath and Uley and the other minor documents collected through the previous work of Amina Kropp (2008), adding new materials coming from the archaeological writing report (Gurney 1986) and publications in the journal Britannia after 2008 (Tomlin 2009; 2010; 2011; 2012; 2013; 2014; 2015; 2016). 36 Francesca Cotugno A Multidisciplinary Analysis of Non-Literary Latin Texts from Roman Britain whereas Vindolanda was garrisoned by Batavians5 and Tungrians in the periods in which the tablets were written (Bowman & Thomas 1994: pp. 22–24). It should be noted that the documents from Carlisle represent the first record of military correspondence writ- ten on tablets from Roman Britain (Tomlin 1998: p. 34). Also, in this corpus, as well as in the others, style6 played a relevant role as the non-classical features collected were not so frequent – at least not as frequent as in the Vindolanda corpus – and they may be interpreted as isolated minor misspellings (e.g. karrissime Tab. Luguv. 40) or as non- classical forms pertaining to a single tablet (i.e. the 16 cases of ordei from Tab. Luguv. 1). Conversely, concerning Vindolanda, the varied records from this corpus are not al- ways easy to interpret. This is the richest set of non-literary records from Roman Britain and the discovery of these wafer-thin tablets in 1973 certainly had an impact on the story of the Roman army and – more importantly for our purposes – on the use of writing, writing material, and Latin by these soldiers (Blackshaw 1974: p. 244). It is noticeable ČLÁNKY / ARTICLES that some of the tablets give clues of a high standard of literacy, maybe due also to the presence of professional scribes: these are documents of praefectorial correspondence, or other formal records such as leave requests, recommendations and memoranda. On the contrary, there are other documents which may be intended as less formal and are more prone to showing non-classical elements. These documents offer a glimpse of the command of spoken Latin of the writers from Vindolanda: these are the personal cor- respondence from people different from the praefects, accounts and lists and other kind of reports (e.g. Tab. Vindol. 181, 343, 344, 595, ex multis). For the sake of clarity, the Vin- dolanda documents may be divided into two main typologies: formal correspondence (composed by letters of recommendation, request of leave, memoranda, reports), and personal correspondence, including those written by women, which can be considered as non-formal. The most recent corpus is composed by the curse tablets (which can be broadly dated between 2nd–4th century CE), which have been largely found in different areas in the southern-western part of this province, mainly at the temples of Bath and Uley (Tomlin 1988; 1993). The writers of these tablets belonged presumably to the local population with a strong influence of Celtic languages (Mullen 2007a: p. 32; Adams 2016: pp. 422, 427; Cotugno, in press). In summary, there seems to be three corpora presumably large- ly written by people coming from outside Britain and one composed by the texts written by people from Britain (Tomlin 1988: pp. 97–98). 5 The 9th cohors of Batavians is the auxiliary cohort which left behind the higher number of writing-tablets at Vindolanda. It was stationed at Vindolanda in the late 80s to early 90s CE. Like all the auxiliary units of the first centuries of this era, it was made up of non-citizen recruits, and came from the region of the Upper Rhine, between the Rhine and the Waal (Battaglia 2013: p. 53). 6 In the sense of the text type involved, which will be briefly considered in § 2.1, and the level of formality of the different documents taken into consideration. 37 Francesca Cotugno A Multidisciplinary Analysis of Non-Literary Latin Texts from Roman Britain 2. Different types of variation The key aim of the analysis of these documents is to show the distribution of the non- classical occurrences of non-literary Latin attested in Roman Britain according to the areal and, when possible, the diachronic axes of linguistic variation. The diffusion of a certain phonological phenomenon and the number of its attestation in a specific area, text type or social strata can help in disambiguating if we are dealing with a mere slip of the pen, an idiolectal form pertaining to a single writer or, conversely, with a well-spread non-classical form among a certain community of writers. As a matter of fact, these different corpora show how Roman Britain had, according to the different community of writers, a differ- ent profile for what concerns at least the phonological variation that can be supposedly be detected in written records. Owing to this, it would be more correct to talk about micro-histories of Latinization instead of the former labels of Latin in Britain and Latin of ČLÁNKY / ARTICLES Britain (Cotugno, in press). Variation among these different communities of writers will be highlighted through the analysis of a few specific and emblematic case studies: vocalic syncope, consonantal gemination and degemination, and h- insertion in initial position. 2.1 Vocalic syncope With reference to vocalic syncope, it is possible to notice that the distribution of this phenomenon differs both geographically and chronologically (see Adamik 2016: pp. 20–21). However, it is relevant to notice that the tablets from Londinium-Bloomberg and Carlisle do not present cases of vocalic syncope (see Table 2). As mentioned above (see § 1), the Londinium and Carlisle corpora belong to a consistently higher level of formality. Conversely, the real comparison is between the Vindolanda corpus and the curse tablets (see Figure 2). Table 2: Vocalic syncope in the non-literary corpora from Roman Britain Corpus Pre-tonic syncope Post-tonic syncope Total Londinium-Bloomberg 0 0 0 Carlisle 0 0 0 Vindolanda 8 (20%) 25 (82%) 33 Curse Tablets 0 3 (100%) 0 Compared to the other corpora (see Table 2), Vindolanda offers a more consistent selec- tion of vocalic syncope.